1: % This is a LaTeX file for a paper by
2: % A.A.Ovchinnikov and M.Ya.Ovchinnikova
3: % A A Овчинников, М.Я.Овчинникова
4: % Figures are accesible as ps-files.
5: %
6: \documentstyle[12pt]{article}
7: \renewcommand{\baselinestretch} {1.35}
8: \topmargin -27pt
9: \textwidth 6.3in
10: \textheight 8.8in
11: %
12: %
13: \def\be{\begin{equation}}
14: \def\ee{\end{equation}}
15: \def\a{\alpha}
16: \def\d{\delta}
17: \def\D{\Delta}
18: \def\g{\gamma}
19: \def\o{\omega}
20: \def\O{\Omega}
21: \def\s{\sigma}
22: \def\G{\Gamma}
23: \def\s{\sigma}
24: \def\e{\epsilon}
25: \def\f{\varphi}
26: \def\z{\zeta}
27: %
28: \begin{document}
29: \vspace{1.5in}
30: \begin{center}
31:
32: {\large \bf Band splitting and relative spin alignment
33: in bilayer systems.} \\
34:
35: \vspace{0.4in}
36: { A.A.~Ovchinnikov$^{1)}$ and M.Ya.~Ovchinnikova$^{2)}$. } \\
37: \vspace{0.2in}
38: {\it $ ^{1,2)}$Joint Institute of Chemical Physics, RAS, Moscow.} \\
39: {\it $ ^{1)}$Max Planck Institute for Physics of Complex Systems, Dresden.}
40: \\ \vspace{0.2in} \end{center}
41:
42:
43: \begin{abstract}
44:
45: Influence of relative spin alignment on the band splitting and magnetic
46: excitations in bilayer correlated systems is studied. Splitting occurs
47: to be large or small depending on relative orientation of staggered
48: spins of the layers. Change of the ground state spin configuration with
49: doping is shown. Behavior of bilayer splitting in $Bi_2Sr_2CaCu_2O_{8+\d}$
50: allows to suppose that superconducting transition is accompanied by
51: transformation of spin configuration of system.
52:
53: \end{abstract}
54:
55: \vspace{0.12in}
56:
57: PACS: 71.10.Fd, 71.27.+a, 71.10.Hf
58:
59: \vspace{0.12in}
60:
61: One of properties of high $T_c$ superconductors - a dependence of their
62: characteristics on number of the $CuO_2$ layers - shows importance of
63: interlayer coupling in them. In recent photoemission experiments \cite{1,2,3}
64: the bilayer splitting of the bands and the Fermi surfaces (FS) in
65: $Bi_2Sr_2CaCu_2O_{8+\d}$ (BSCCO) have been observed. Unusual observation was
66: a significant decrease in bilayer splitting at $k\sim (\pi,0)$ in course of
67: the superconducting (SC) transition \cite{2}. Another phenomena which is
68: observed in the bilayer (but not monolayer) cuprates is the appearance in
69: SC state of a spin resonance in inelastic neutron scattering
70: \cite{4,5,6}.
71: Explanation of these phenomena is important for understanding the electronic
72: structure and mechanisms of HTSC displaying many properties of doped Mott
73: insulator. The main feature of latter is the Hubbard splitting of band
74: induced by antiferromagnet (AF) fluctuations. Though the long range AF order
75: disappear at small doping $\d_c=1-n_c\sim 0.05$, the local 2D spin order
76: probably retains in sufficiently large doping range including the range of
77: superconductivity. Among arguments in favour of such local AF order
78: there are data on $\mu$SR, NMR, NQR \cite{7,8}, large length of spin
79: correlations with $Q=(\pi ,\pi )$ and smooth evolution of collective magnetic
80: excitations with doping (see review \cite{4}) and others. Finally, recently
81: a direct proof of the AF order even in SC state of
82: $YBa_2Cu_3 O_{6.5}$ have been obtained \cite{9,10} in nanosecond scale.
83:
84: Detailed analysis of the AF bands and interactions of the Hubbard model have
85: been done in classic work \cite{11}. Simple mean field method gives
86: the overestimated values $d_0$ of staggered spin and of the doping boundary
87: $\d_c\sim 0.45$ of a range of the AF order. The band renormalization on base
88: of the zero AF approximation \cite{11} or the calculations of t-t'-U
89: Hubbard model using the slave-boson technique \cite{12} or the states with the
90: valence bond correlations \cite{13} - all these methods give the smaller
91: values of $d_0$ and of the boundary doping
92: $\d_c\sim0.3$ at which the local magnetization disappears. According
93: \cite {13} the range $\d<\d_c$ includes entirely the range of
94: superconductivity. Retaining of the Hubbard splitting of band at $\d<\d_c$
95: in the t-t'-U \cite{13} and t-t'-J \cite{14,15,16} models implies the change
96: of the FS topology at the optimal doping with transition from the hole
97: pockets to large FS.
98:
99: If such picture is true then interactions between the layers in bilayer
100: systems may significantly depend on relative orientation (alignment) of local
101: staggered spins of the layers. It must be so even if the spin orientation is
102: not stationary in strong sense, but is retained only on some time interval.
103: In case of small difference in the energies of various configuration of spin
104: alignment one can expect a strong influence of doping and temperature on
105: the properties of magnetic excitations in system. A search of such effects
106: may help to elucidate the origin of the magnetic resonance in neutron
107: scattering in YBCO, BSCCO.
108:
109: Present work summarizes the results of variational calculations of bands,
110: full energies, phase curves $T_c(\d )$ and the band splittings in the
111: bilayer t-t'-U Hubbard model. Calculations are made on base of the correlated
112: states with the valence bond (VB) formation - the band analog of the RVB
113: state of Anderson. Previously \cite{13} it was shown that the correlated
114: hopping interaction appearing in the effective Hamiltonian during the VB
115: formation does provide the hole attraction in d- channel and corresponding
116: d-superconductivity compatible with AF spin order. Here we use the same
117: method for study the effects of interactions of two layers. The questions of
118: interest are: Is the large doping range of local AF order retained? If so,
119: then what are the relative alignment of spin systems of both layers and the
120: bilayer splitting at various doping?
121:
122: A difference between given approach and the study \cite{17} of bilayer
123: Hubbard model by the FLEX (fluctuation exchange) method must be stressed.
124: According FLEX the hole attraction is provided by long range spin
125: fluctuations. In our approach the short range AF correlations are decisive.
126: In simpliest variant they correspond to the VB formation, i.e. formation of
127: singlets on the bonds of the nearest centers. Unlike the averaged
128: consideration in FLEX we make the calculations for two specific spin
129: configurations of bilayer system - homogeneous states with the same or
130: opposite alignment of staggered spins of both layers. We denote them as the
131: $F_z$ or $AF_z$ (ferro- or antiferro-magnet over z direction) alignment
132: correspondingly, though both cases refer to local AF spin order in each
133: layer. We consider only homogeneous states, though many neutron experiments
134: give indication on the stripe-type or spiral incommensurate structures in
135: cuprates. Nevertheless the effects discovered in our calculations for two
136: homogeneous spin structures must undoubtedly be taken into account in
137: discussing the observed bilayer splitting and magnetic properties of cuprates.
138:
139: We start from the Hubbard Hamiltonian of bilaeyr system $H(U,t,t',t_z)$.
140: The main intralayer interactions are determined by
141: standard parameters $U,t$ of Hubbard model. Additional interactions include
142: the hoppings $t', ~t_z$ between the next nearest centers inside the layers
143: and between the cites of different layers. Positive signs of standard
144: parameters of strong coupling $t,~t',~t_z$ are defined in such way that the
145: zero bands (at $U=0$) have a form
146: \be
147: \e_{\pm}^{(0)}=-2t(cos{k_x}+cos{k_y}) +4t'cos{k_x}cos{k_y}
148: \mp {1\over 4}t_z(cos{k_x}-cos{k_y})^2
149: \label{1}
150: \ee
151: Expression for splitting $\D\e (k)=\e_+ -\e_- $ between the bonding and
152: antibonding zero bands is derived in \cite{17}.
153:
154: A variational correlated state $\Psi=\hat{W}(\a)\Phi$ with correlations of
155: the VB type is constructed \cite{13} as unitary transformed uncorrelated
156: state $\Phi$
157: \be
158: \hat{W}(\a)=\exp\Bigl[\a Z\Bigr]; \quad Z= {{{1\over 2}}}
159: \sum_{\s,<nm>,\g}(c_{\g n\s}^\dagger c_{\g m\s}-h.c.) (n_{\g n,-\s}-n_{\g
160: m,-\s});
161: \label{2}
162: \ee
163: Here $\g=0,1$ refer to layers a,b.
164: A choice of unitary operator $W(\a)$ with variational
165: parameter $\a$ is explained in \cite{13}. Variational treatment of
166: problem with Hamiltonian $H$ in basis of correlated state $\Psi$ is
167: equivalent to treatment of the effective Hamiltonian
168: $H_{eff}(\a)=W^{\dagger}(\a)HW(\a)\approx H +\a [H, Z]
169: + {\a^2\over 2} [[H, Z]$
170: in mean field approximation. We use the most general uncorrelated states
171: $\{\Phi\}$ of the BCS type with anomalous averages of d-symmetry and the AF
172: spin order. The mean energy over such state
173: $<H>_{\Psi}=<H_{eff}>_{\Phi}=\bar{H}(y_{\g\nu},z)$
174: is obtained as an explicit function of a set of one-particle averages
175: $\{y_\nu , z\}$ over uncorrelated state $\Phi$. The set of these averages
176: contains the density components $r_{\g l}=<c_{\g n\s}^\dagger c_{\g,n+l,\s}>$,
177: analogous components of staggered spin
178: $d_{\g l}=(-1)^n {\s\over \mid\s\mid}<c_{\g n\s}^\dagger
179: c_{\g ,n+l,\s}>$, anamalous averages
180: $w_{\g l}=(l_x^2-l_y^2){<c_{\g n\s} c_{\g,n+l,\s}>}$ of d-symmetry for each
181: layer $\g =0,1$ and a value $z=t_z^{-1}<T_z>$ determining the average
182: interlayer hopping. For spin components we study two variants of relative
183: alignment $d_{\g =1,l}=\zeta_d d_{\g=0,l},~~\zeta_d=\pm 1$.
184: Selfconsistent procedure of the energy minimization over $\Phi$ and
185: of subsequent minimization over $\a$ is standard one \cite{13}. Details of
186: calculations will be presented elsewhere.
187:
188: For main parameter we use a value $U/t=8$ following from
189: cluster derivation of single band model. Parameter $t'$ owing to its
190: contribution into the band dispertion regulates the position
191: $E_{VHS}=E(\pi,0)$ of the van-Hove singularity (VHS) in the density of
192: state relative to the edge of the lower Hubbard band. In the monolayer
193: $t-t'-U$ or $t-t'-J$ models a value of $t'$ is directly connected with the
194: optimal doping $\d_{opt}$, since a maximum of $T_c$ corresponds to
195: coincidence of the chemical potential $\mu$ with $E_{VHS}$. According the
196: models \cite{13,14,15,16} a resonable value of
197: $\d_{opt}\sim 0.2\div 0.25$ corresponds to values $t'/t\sim 0.05\div 0.1$.
198: These values are less than $t'/t\sim 0.2\div 0.4$ obtained from
199: fitting of photoemission data and the band calculation data on base of
200: the strong coupling scheme. Here we vary parameter $t'$ in
201: limits $t'=0.05 \div 0.3$. Parameter of interlayer hopping is also varied
202: as $t_z/t=0.07\div 0.3$. Previous calculations without renormalization gave
203: $t_z/t\sim 1/3$ \cite{17}. Recent measurements of bilayer splitting
204: $\D=2t_z=\d\e_k(\pi,0)$ \cite{1,2} give the value $t_z/t\leq 0.1$.
205:
206: Fig.~1 presents typical doping dependencies of spin density $d_0$,
207: transition temperature $T_c(\d)$ and a difference
208: $\D H={\bar H}(AF_z)-{\bar H}(F_z)$ of average energies of bilayer system
209: with two types of relative alignment of the layer's spins. Staggered spin
210: density $d_0$ has nonzero value at wide range of doping including the region
211: of supercondutivity. Quantity $\D H(\d)$ changes its sign at some doping. In
212: the undoped system, $\d=0$, the
213: configuration $AF_z$ with opposite alignment of the
214: layer's spins occurs to be lower than the $F_z$ configuration. At large
215: doping $\d$ one have inverse situation up to boundary value $\d_c$, at
216: which the local magnitization disappears. Negative sign of $\D H$ at
217: $\d=0$ is explained by a positive exchange constant $J_{ab}\sim 2t_z^2/U>0$.
218: Change of sign of $\D H$ at some $\d$ is a result of the bilayer splitting of
219: the bonding and antibonding bands, in particular, the splitting of VHS, and
220: different occcupancy of them for two types of spin alignment.
221: Maximum of $\D H(\d )$ corresponds to optimal doping $\d_{opt}$
222: for the parent monolayer model. At this doping one has
223: $E_{VHS}^{1L}-\mu=0$. Increase of $t'$ shifts both the $\d_{opt}$
224: and the position of maximum of $\D H(\d )$. Such relation is not occasional.
225:
226: According \cite{17} the bilayer splitting
227: $\d\e^{(0)}_z=2t_z(\cos{k_x}-\cos{k_y})^2/4$
228: of the zero bands has a maximum in region $k\sim (\pi ,0)$ forming the VHS
229: in density of state. However, the splitting contributes to the average
230: energy only if at this $k$ bonding and antibonding bands appear to be
231: occupied and unoccupied correspondingly. The latter takes place when
232: $E^{1L}(\pi,0)=E_{VHS}^{1L}=\mu$ for the unsplit bands of the monolayer model.
233: It remains to explain why a bilayer splitting manifests itself in the energy
234: difference $\D H={\bar H}(AF_z)-{\bar H}(F_z)$. The reason is in different
235: bilayer splitting for different spin configurations.
236:
237: Fig.~2 presents the density of state of lower Hubbard band for two
238: configurations of spin alignment for system with small values $t',~t_z$.
239: Bilayer splitting is absent at $AF_z$ alignment, but sharply seen in $F_z$
240: configuration. This is due to different behaviour of matrix elements of
241: the interlayer hopping between the states of lower Hubbard band of each layer.
242: At $AF_z$ alignment these matrix elements disappear on the nesting lines.
243: Different behaviour of DOS reflects itself directly on the form of the phase
244: curves $T_c(\d)$ for SC transition (Fig.~1). In case of $AF_z$ or $F_z$
245: configuration the curve $T_c(\d)$ has one or two maxima correspondingly.
246:
247: Thus the models with small $t',~t_z\mathop{_\sim^<} 0.1$ are characterised
248: by 1) large $T_c^{max}\sim 0.02t\sim 116K$ at $AF_z$ configuration; 2) close
249: energies of both configurations, 3) small condensation energy; 4) lower $F_z$
250: configuration at $\d\sim \d_{opt}$; 5) large (or zero) bilayer splitting
251: in $F_z$ (or $AF_z$) configuration.
252:
253: Unlike the normal state, a prediction of a lower spin configuration in
254: the SC state is hardly possible. An estimate of the condensation
255: energy and the step in a heat capacity occurs to be less than that observed
256: in YBCO approximately in 5 times. For this reason we dicuss the spin
257: configuration of SC state on base of the observed behaviour of bilayer
258: splitting in BSCCO \cite{1,2}. This splitting decreases from $\d\e\sim 80
259: meV$ in normal state down to $\d\e\sim 20 meV$ in SC state. Such behaviour
260: might be explained if we suppose that a transformation of configuration
261: $F_z\to AF_z$ proceeds simultaneously with SC transition. One might suppose
262: also that the SC transition as such is induced by the change of
263: configuration, since such transition is accompanied by increase of the
264: density of states at the Fermi level.
265:
266: For models with large values $t',t_z$ the types of dependencies $\D H(\d )$,
267: $d_0 (\d )$ are retained. Increase of $t_z$ leads to increase of $|\D H|$ and the doping
268: of the $F_z~-AF_z$ crossover. Calculations confirm the wide doping range of
269: AF local spin order for these model also. Value of $d_0$ is greater by an
270: order of magnitude than the value of AF spin density observed in SC state of
271: YBCO \cite{9,10} from elastic neutron scattering. The difference, possibly,
272: is connected with large distribution of the spin directions (or signs) of
273: different biplaines in crystal. Significant decrease of $T_c$ and a
274: deformation of the phase curves with increase of $t',~ t_z$ are connected
275: with decrease of the density of state on the FS for these models.
276:
277: Influence of the spin alignment on the magnetic excitation spectrum
278: is of interest in connection with discovery of the spin resonance in neutron
279: scattering in SC state of cuprates and its unusual dispersion \cite{10}.
280: A widely discussed hypothesis \cite{18} connects its origin with the so
281: called $\pi$-resonance - an excitation of $e-e$ pair with $Q\sim (\pi,\pi )$
282: in triplet state. But the ratio of weights of such resonance to the integral
283: (over $\o$) intensity of the non-resonance
284: magnetic excitations would be too small.
285: Expected order of magnitude of this ratio is $\sim (w/d_0)^2\sim 0.01$.
286: Interpretation of incommensurate patterns in the spin
287: susceptibility on base of the
288: nesting properties of FS also cannot describe a large intensity of spin
289: excitations and their similarity in both the undoped dielectric materials
290: and the doped metallic ones. The most probable hypothesis \cite{19} implies
291: the common magnetic origin of spin resonance and of the incommensurate
292: anomalies of spin susceptibility $\chi ''(q,\o )$ in cuprates. Both features
293: are explained by existence of the AF domains or the incommensurate modulation
294: of the local staggered magnetization. The stripe-type structures have been
295: observed in $LaSrCuO$ and they may also present in the bilayer cuprates. But
296: monolayer models \cite{19} cannot explain the fact that the resonance
297: appears only in superconducting state and only in odd channel of bilayer
298: system. In this connection it is important to study the influence of spin
299: alignment on magnetic excitations in bilayer systems.
300:
301: Here we study such effects using a simpliest example of homogeneous model
302: for "constructing" the spin resonace. Consider the fenomenological spin
303: Hamiltonian with anisotropic interlayer interaction
304: \be
305: H=\sum_{\g =a,b}J_0\sum_{<nm>}{\bf S}^{\g}_ n{\bf S}^{\g}_ m+
306: \sum_{n}\{ J_{z,ab}{S}_{zn}^a S_{zn}^b
307: +J_{\perp,ab}[S_{xn}^a S_{xn}^b+S_{yn}^a S_{yn}^b ]\}
308: \label{3}
309: \ee
310: %\be
311: %V_{ab}=\sum_{n}\{ J_{z,ab}{S}_{zn}^a S_{zn}^b
312: %+J_{\perp,ab}[S_{xn}^a S_{xn}^b+S_{yn}^a S_{yn}^b ]\}
313: %\label{4}
314: %\ee
315: The "$z$"-direction is selected by vector of local staggered spins $<{\bf
316: S}^\g_n>={\bf e}_z (\z_d)^\g d_0$ of each layer ($\g=0,~1$ for layers a,b)
317: with definite alignment $\z_d=1$ or $-1$, stabilized by the nonspin
318: interactions. Then the RPA consideration in the linear spin wave theory
319: \cite{20} gives the following frequencies of collective spin excitations
320: for both types of alignment
321: \be
322: \begin{array}{lll}
323: F_z:&~~~\o^{even}(q)\approx 2\O\sqrt{\f-g_1},&~~
324: \o^{odd}(q)\approx 2\O\sqrt{\f -g_2};~~~~\\
325: AF_z:&~~~\o^{even}(q)\approx 2\O\sqrt{\f+g_2},&~~
326: \o^{odd}(q)\approx 2\O\sqrt{\f +g_1};~~~ \\
327: \end{array}
328: \label{5}
329: \ee
330: Here $\f=2+\cos{q_x}+\cos{q_y}$, $\O=2J_0d_0$,
331: $g_{1(2)}=(J_{z,ab}\mp J_{\perp , ab})/2J_0$.
332:
333: Fig.~3 presents a dispersion of the spin excitations in even and odd channels
334: for both configurations of a spin alignment for model with parameters
335: $g_1=0.05$, $g_2 =0.15$. Consider them in light of the above suggestion
336: about simultaneous transformation $F_z\to AF_z$ of the spin configuration
337: and superconducting transition in bilayer cuprates. If it is true for our
338: spin model (3), then at $T<T_c$ the intense peak appears in the odd channel
339: of $\chi^{odd}$ at $q\sim (\pi,\pi)$ with the frequency
340: $\o=\D^{odd}=2\O\sqrt{g_1}$, equal to a gap in the excitation spectrum in this
341: channel for $AF_z$ spin configuration. This peak is absent in case of $F_z$
342: configuration which is supposed to be ground state configuration at $T>T_c$.
343: Note that for $F_z$ configuration the low frequency excitations in
344: $\chi ''(q,\o\to 0)$ correspond to momenta $q$ distributed along the circle
345: $|{\bf q}-{\bf Q}|=\sqrt{2g_1}$ instead of the discrete incommensurate
346: momenta $Q_{\eta}=(\pi (1+\eta ),\pi )$, observed for the low frequency
347: excitations in YBCO \cite{13}.They imply an existence of the incommensurate
348: spin structures. Fig.~3c schematically presents the branches of the spin
349: excitations for the spiral states or the states with modulation of a local
350: spin {$<S_{zn}>=d_0\cos{{\bf Q}_\eta {\bf n}}$}
351: of the monolayer (1L) model \cite{20}.
352: In \cite{20} the resonance frequency have been identified as a frequency at
353: the crossing point of branches at $q=(\pi ,\pi )$.
354:
355: Thus, it is confirmed that local magnitization $d_0$ is retained in large
356: doping range for wide diapason of parameters $t',~t_z$. The crossover of two
357: spin configurations of the bilayer system proceeds at some doping. The lower
358: normal state is that with the $AF_z$ or $F_z$ spin alignment at small or large
359: doping correspondingly. Maximum of the energy difference of these
360: configurations corresponds to the optimal doping of the parent monolayer
361: system and is connected with the maximum splitting of van-Hove singularities.
362: The bilayer splitting is zero or small for the $AF_z$ configuration, but
363: large for the $F_z$ spin alignment. The observed in BSCCO large bilayer
364: splitting in normal state, but small splitting in SC state \cite{2} may
365: indicate according our models that the transformation $F_z\to AF_z$ of
366: configuration proceeds simutaneously with supercunducting transition. It is
367: shown that the magnetic excitation spectrum depends dramatically on type of
368: a spin configuration. The model spin system is presented for which a change
369: of configuration $F_z\to AF_z$ is accompanied by appearance of resonance (or
370: spin gap) in $\chi^{odd}$ at $q=(\pi,\pi )$.
371:
372: A check of the above hypothesis requires further study, in particular
373: calculations of magnetic excitations in case of the inhomogeneous (of the
374: stripe type) spin structures of bilayer systems.
375:
376: Work is supported by Russian Fund of Fundamental Reasearch (Projects
377: No. 00-03-32981 and No. 00-15-97334. Athours thanks V.Ya Krivnov for useful
378: discussions and P.Fulde for possibility to work in
379: Max Planck Institute for Physics of Complex Systems, Dresden.
380:
381: \vspace{0.2in}
382:
383: %\newpage
384: \begin{thebibliography}{99}
385:
386: \bibitem{1}
387: Y.-D.Chuan, A.D.Gromko, A.Fedorov et al. E-print archiv
388: cond-mat/0102386
389:
390: \bibitem{2}
391: D.L.Feng, N.P.Armitage, D.H.Lu et al. E-print archiv
392: cond-mat/0102385
393:
394: \bibitem{3}
395: J.Mesot, M.Boehm, M.R.Norman et al. E-print archiv cond-mat/0102339
396:
397: \bibitem{4}
398: P.Bourges in "The gap Symmetry and Fluctuations in High Temperature Superconductors",
399: ed. J.Bok, G.Deutscher, D.Pavuna, S.A.Wolf (Plenum Press).
400:
401: \bibitem{5}
402: P.Bourges, Y.Sidis, H.F.Fong et al. Science {\bf 288}, 1234 (2000).
403:
404: \bibitem{6}
405: H.A.Mook, P.Dai, F.Dogan, R.D.Hunt et al. Nature {\bf 404 }, 729 (2000).
406:
407: \bibitem{7}
408: A.Rigamonti, F.Borza, P.Carett, Rep.Progr.Phys. {\bf 61}, 1367 (1998).
409:
410: \bibitem{8}
411: Ch.Niedermayer, C. Bernhard, T.Blasius et al. Phys.Rev.Lett. {\bf 80},
412: 3843 (1998).
413:
414: \bibitem{9}
415: H.A.Mook, P.Dai, F.Dogan, R.D.Hunt et al. E-print archiv cond-mat/0102047.
416:
417: \bibitem{10}
418: Y.Sidis, C.Ulrich, P.Bourges et al. Phys.Rev.Lett. {\bf 86} 4105 (2001).
419:
420: \bibitem{11}
421: J.R.Schrieffer, X.G.Wen, F.C.Zhang, Phys.Rev. {\bf B 39}, 11663 (1989).
422:
423: \bibitem{12}
424: U.Trapper, D.Ihle, H.Fenke, Phys.Rev. {\bf B 52}, 11553 (1995).
425:
426: \bibitem{13}
427: A.A.Ovchinnikov, M.Ya.Ovchinnikova, Phys.Lett. {\bf A249}, 531 (1998).
428: A.A.Ovchinnikov, M.Ya.Ovchinnikova, E.A.Plekhanov, JETP Letters,{\bf 67},
429: 350 (1998); JETP {\bf 114}, 985 (1998); {\bf 115}, 649 (1999)].
430:
431: \bibitem{14}
432: N.M.Plakida, V.S.Oudovenko, R.Horsch, and A.J.Liechtenstein, Phys.Rev. {\bf
433: B55}, 11997 (1997).
434:
435: \bibitem{15}
436: R.O.Kuzian, R.Hayn, A.F.Barabanov, and L.A.Maksimov, Phys.Rev. {\bf B 58},
437: 6194 (1998).
438:
439: \bibitem{16}
440: F.Onufrieva, P.Pfeuty, and M.Kisilev, Phys.Rev.Lett {\bf B 82}, 2370 (1999).
441:
442: \bibitem{17}
443: A.I.Liechtenstein, O.Gunnarsson, O.K.Andersen, R.M.Martin
444: Phys.Rev. {\bf B 54}, 12505 (1996).
445:
446: \bibitem{18}
447: E.Demler, S.C.Zhang Phys.Rev.Lett. {\bf 75}, 4126 (1995).
448:
449: \bibitem{19}
450: C.D.Batista, G.Ortitz and A.V.Balatsky, E-print archiv cond-mat/0008345.
451:
452: \bibitem{20}
453: T.Ziman, P.-A.Lindgard, Phys.Rev. {\bf B 33}, 1976 (1986).
454:
455: \end{thebibliography}
456: \vspace{0.2in}
457:
458: %\newpage
459: {\bf Captions to Figures}
460: \vspace {0.15in}
461:
462: Fig.~1.
463: Doping dependence of the local magnitization $d_0(\d )$ (top), of the
464: $T_c(\d )$ (middle) and value $w_1(\d)$ of the anamalous order parameter
465: (bottom) for configurations $AF_z$ и $F_z$ (curves 1 and 2 correspondingly).
466: Curve 3 is the difference of energies of these configurations. All
467: values $T_c$, $t_z,~t'$, $\D H$ are in unit $t$.
468:
469: Fig.~2.
470: Density of state for two types $AF_z$ and $F_z$ of spin alignment
471: for models with small $t_z, t'$. Thin solid and dashed lines refer to
472: contributions in DOS from bonding and antibonding bands. The calculated DOS
473: is smoothed with gaussian function with dispersion $\d E=0.02t$, E is in
474: unit $t$.
475:
476: Fig.~3.
477: Changes in the magnetic excitations branches $\o(q)$ during transformation
478: $F_z\to AF_z$ of spin configuration of model (5) with anisotropic
479: interlayer interaction with parameters $g_1=0.05$, $g_{2}=0.15$.
480: A right figure is schematic presentation of the spin excitations in case of
481: the spiral or modulated incommesurate spin structure with
482: $Q_{\eta} =(\pi (1-\eta ), \pi )$, $\eta=0.1$.
483: Squares mark the supposed resonance frequencies in various
484: interpretations.
485:
486:
487:
488: \end{document}
489: